How to deploy Azure ML batch endpoint from docker image?

Question

Hi, I have my own deep learning task that requires 2-3 different ml models, I built the code and containerized it, i.e. the python env and code is in the docker image.

I am running fastapi servers inside docker to run code.

Deployed it in aws sagemaker async endpoint and it is working fine.

Now, I need to deploy it to azure ml batch endpoint, but there's no documentation as such to deploy it using custom docker container.

Can someone help me?

Answer

Deploying a custom Docker container to an Azure ML batch endpoint can be tricky due to limited documentation. However, see if the steps below will help you:

Prepare Your Docker Image: Ensure your Docker image is ready and pushed to a container registry accessible by Azure (e.g., Azure Container Registry).

Create an Azure ML Workspace: If you haven’t already, set up an Azure ML workspace.

Set Up Your Environment: Install the Azure CLI and the Azure ML extension:
```
   az extension add -n ml
```
Create a Batch Endpoint: Use the Azure CLI or Python SDK to create a batch endpoint. Here’s an example using the CLI:
```
   az ml batch-endpoint create --name  --file 
```
Deploy Your Model: Create a deployment for your batch endpoint using your custom Docker image. Here’s an example configuration:
```
   name:
```
Test Your Deployment: Once deployed, you can test your batch endpoint to ensure everything works as expected.

References:

https://video2.skills-academy.com/en-us/azure/machine-learning/how-to-deploy-custom-container?view=azureml-api-2&tabs=cli

https://video2.skills-academy.com/en-us/azure/machine-learning/how-to-use-batch-model-deployments?view=azureml-api-2&tabs=cli

Please remember to "Accept the answer” and “up-vote” wherever the information provided helps you; this can benefit other community members.

Share via

How to deploy Azure ML batch endpoint from docker image?

1 answer

Your answer